Building a Context Image-Based Search Engine Using Multi Clustering Technique

Content-Based Image Retrieval (CBIR) is a challenging task which retrieves the similar images from the large database. Most of the CBIR system uses the low-level features such as color, texture and shape which has not much detailed information about the images, in case of looking for images that contain the same object or same scene with different viewpoints to extract the features from the images. Allowing users to find images on the web similar to a particular query image is a crucial component of modern search engines. In this paper the Speeded Up Robust Feature is combined with the color feature to improve the retrieval accuracy of the search engine. In this paper, k-means clustering algorithm is used for clustering image features. However, it is computationally expensive and the quality of the resulting clusters heavily depends on the dimension of the data. This paper proposed a new approach to improve the accuracy of the cluster results from using a new novel algorithm called NCD to reduce the dimension of the image features in the dataset. Experiment results show that the proposed color feature is more accurate and efficient in retrieving images with user-interested color and image objects compared with the current algorithms. Speeded Up Robust Features (SURF) show its advantages in rotation, scale changes, image blur, affine transformations and illumination changes.